Referring to FIG. 1, DWDM networks 20 implemented over fiber rings 22 can carry diverse types of traffic such as, for example, SONET, ATM, IP, and so on. These networks 20 are capable of mixing different types of traffic in the same ring 22. A typical DWDM network 20 includes an arbitrary number of nodes 24 interconnected in a ring topology by a pair of optical fibers 26. One of the nodes 24 is designated the hub node 30. The other nodes 24 are referred to as terminal nodes 32. Each terminal node 32 uses one or more dedicated DWDM wavelengths λJ, λK, . . . λP, 1≦J, K, . . . P≦N, to communicate with the hub node 30. The hub node 30 has the capability to switch traffic from one wavelength λ1, λ2, . . . λN to another. This permits communication between any pair of terminal nodes 32 on the network 20. The DWDM channel λ1, λ2, . . . λN used to transmit traffic from the hub node 30 to a specific terminal node 32 over one of the fibers 26 is called a downlink. The DWDM channel λ1, λ2, . . . λN of the same wavelength operating on the other fiber 26 used to transmit traffic from the terminal node 32 to the hub node 30 is called an uplink. The resulting network 20 is sometimes described as a virtual DWDM star network implemented over a fiber ring 22. The protocol used in the interaction between the hub node 30 and a specific terminal node 32 is arbitrary and independent of the protocol used by any other terminal node 32. Examples of protocols include the above-mentioned SONET/SDH, ATM and IP. Where different channels λ1, λ2, . . . λN use different protocols, all channels λ1, λ2, . . . λN may be assumed to be using a common protocol, for example, SONET/SDH framing, with the other protocols, for example, ATM, IP and so on, mapped into the assumed common protocol (SONET/SDH frames in this example). Both the hub node 30 and the terminal nodes 32 have the capability to effect the appropriate protocol processing on both incoming and outgoing traffic. All nodes 24, including the hub 30, have local tributary interfaces which permit the connection of external equipment to the network 20.
A network 20 as describe above is expected to be extremely reliable and remain fully or at least partially operational despite faults of different types. Of special, although not exclusive, interest in the context of this application are the following types of faults: the failure of a transceiver in one of the nodes 24; a break or other malfunction in the physical fiber 26 that renders a segment of the ring 22 unusable; and, total or partial failure of the hub node 30.